Genetically predicted blood pressure, antihypertensive drugs and risk of heart failure: a Mendelian randomization study.

Abstract

Elevated blood pressure (BP) was associated with higher risk of heart failure, but the relationship between BP-lowering via antihypertensive drugs and diminution of heart failure was inconclusive. This study aimed to estimate the causal association of BP with heart failure, and explore the effects of BP-lowering through different antihypertensive drug classes on heart failure risk using Mendelian randomization analysis with genetic variants as instrument variables. Genetic variants associated with BP were derived from UK Biobank ( n = 317 754) and the genome-wide association study (GWAS) meta-analysis of UK Biobank and International Consortium of Blood Pressure ( n = 757 601). Heart failure summary association data were contributed by HERMES Consortium (47 309 heart failure cases and 930 014 controls). Inverse variance weighted (IVW) was performed to estimate causality between exposure and outcome, and weighted median was utilized as sensitivity analysis, and Mendelian randomization-Egger regression was used to identify pleiotropy of instrument variables. Multivariable Mendelian randomization (MVMR) was applied to control for the confounders. Genetically predicted SBP and DBP were associated with heart failure [SBP: odds ratio (OR) = 1.355, 95% confidence interval (CI) 1.201-1.529; DBP: OR = 1.348, 95% CI 1.213-1.498] in UK Biobank. Likewise, in the GWAS meta-analysis of UK Biobank and International Consortium of Blood Pressure, the causal associations were observed between SBP, DBP and heart failure (SBP: OR = 1.237, 95% CI 1.188-1.289; DBP: OR = 1.337, 95% CI 1.245-1.437). Genetically determined β-blockers and calcium channel blockers (CCBs) were associated with lower risk of heart failure (β-blockers: OR = 0.617, 95% CI 0.453-0.839; CCBs: OR = 0.730, 95% CI 0.625-0.851). No association was found between angiotensin receptor blockers (ARBs) and heart failure (OR = 1.593, 95% CI 0.647-3.924). When adjusted for smoking, alcohol, physical activity, fruit and vegetable intake, the results were stable. Our study indicates causal associations between SBP, DBP, and heart failure, and suggests the preventive effects of heart failure by BP-lowering using β-blockers and CCBs.